The goal of this project is to develop recombinant vaccines which can activate antitumor T cells to prevent the growth or recurrence of breast cancer. The investigators will test the hypothesis that recombinant tumor associated antigens targeted to the major histocompatibility complexes (MHC) I and II antigen processing pathways have enhanced presentation and are strong immunogens. ERBB-2 will be used to test the principle and to generate potential cancer vaccines. A truncated ERBB-2 (cyt ERBB-2) construct without ER signal peptide sequence has been generated. Cells transfected with the pcDNA/cyt ERBB-2 express high levels of cytoplasmic ERBB-2 which is conjugated to ubiquitin and is promptly degraded by the proteasome and may be a rich source of MHC I associated antigenic peptides. The kinase activity of cyt ERBB-2 which is associated with transforming activity will be eliminated by replacing ATP binding lysine residue 753 with alanine. To direct ERBB-2 into MHC II processing pathway, mutant ERBB-2 (lys ERBB-2) will be generated by replacing the cytoplasmic domain of ERBB-2 with that of lysosomal associated membrane protein (LAMP-1). The immunogenicity of wild type, cyt, and lys ERBB-2 will be tested with both human and murine cells. Dendritic cells have been generated with human and mouse CD34+ cells as antigen presenting cells. The vaccination efficacy will be tested in mice with naked plasmid DNA, adenoviral vectors, and gene modified antigen presenting cells. Specifically, the investigators will A) continue to modify and test recombinant vectors encoding antikinase or nonfunctional ERBB-2 antigenic presentation by MHC I and II, B) test the activation of T cells in induction of antitumor immunity with recombinant ERBB-2 vaccines. Results from the study will guide future clinical trials for patients entering KCI breast cancer bone marrow transplant program.